On the Relations between VU-Meter Readings and the rms and Peak Values of Speech
Weiant Wathen‐Dunn · The Journal of the Acoustical Society of America · 1961
In articulation testing, the signal-to-noise ratio is very often one of the parameters against which one wishes to plot the articulation score, and ideally this S/N should be the ratio between the long-time rms value of the signal and a similar measure of the noise. When speech is the signal, however, it is difficult, if not impossible, to determine its true rms value with any great accuracy by means of the instruments usually available unless one has established relationships between the readings of these instruments and the desired measure of the signal which are valid for the particular conditions of the test. Since the VU meter is most often used to monitor the level of the speech in articulation test work, it is particularly desirable to know the relation between its readings and the true rms value of the signal. An alternative method of arriving at this figure would be to take peak-value readings of the signal from an oscilloscope from which the rms value could be calculated provided the peak factor were known. The present paper is concerned with measuring the relation between the average value of peak VU-meter readings and the rms and peak values of both wide- and narrow-band samples of continuous speech recorded by several adult male and female talkers.